1. Field of the Invention
The present invention relates to a peer-to-peer (P2P) type communication system including a plurality of node devices mutually communicable through a network.
2. Discussion of Related Art
As this kind of system, there is known a technique of so-called application layer multicast (ALM) which relays to deliver a content stream while broadcast (deliver) routes of the content data are managed at an application layer. In this ALM, packets generated from content data files are relayed in real time between node devices, thereby delivering content data broadcasted from a broadcast station (broadcast device) to arbitrary node devices (multicast).
In a case of reconnection to a new upper hierarchy node device triggered by quality deterioration of receiving stream (content stream from the node device of the upper hierarchy) or withdrawal (separation) of the upper hierarchy (upstream) node device, the node device participating in a topology formed under such ALM uses a method where the node device requests a connection destination introduction server managing topology to thereby acquire network address information of a connection destination candidate and uses this to reconnect, or a method that the node device autonomously reconnects in use of network address information of the connection destination candidate which respective node devices exchange with each other and store. It is often the case that either of the above methods is used.
By the former method, because an appropriate connection destination candidate can be introduced from entire topology information collected in real time by the connection destination introduction server, it is possible to carry out a robust connection process in use of the connection destination candidate. However, since burden loaded on the connection destination introduction server increases in a case where a lot of node devices participate in the topology, there is a problem that the system tends to be unstable in its entirety.
By the latter method, since reconnection is carried out in use of network address information that respective node devises exchange with neighbor node devices, burden is not loaded on the connection destination introduction server for acquiring the connection destination candidate. However, since it is difficult to select an appropriate connection destination candidate from the entire topology, there is a high possibility that a connection trial on a node device to be a connection destination candidate is failed to necessitate retrial of connection.
Further, Patent Document 1 discloses that a connection destination introduction server manages a predetermined number of connection destination candidates with respect to every attribute information of the node devices, and the node device which desires participation in a topology and accompanies the own attribute information requests the connection destination introduction server to thereby acquire a connection destination candidate. By this method, the node device requests a node device to be connected for connection in use of network address information of the connection destination candidate which is acquired from the connection destination introduction server in response to the connection destination candidate acquisition request, and the connection is established in a case where the node device to be connected permits connection. According to this method, the appropriate connection destination candidate is not selected in the connection destination introduction server but the node device to be connected responds whether or not the connection is permitted, after judging conditions of the node device requesting connection in detail. Therefore, it is possible to reduce a load related to the process for the connection destination introduction server to select the appropriate candidate.
[Patent Document 1] Japanese Unexamined Patent Publication No. 2006-319466